The invention relates to any type of rotating machinery such as centrifugal pumps, water and gas turbines, electric motors and generators. The invention deals with the way to reduce the frictional drag that is present in this types of machinery due to the rotating components of the machinery.
It is inevitable that the rotating components on such machinery will experience a retarding drag force that is developed by the fluid surrounding the rotating component. To facilitate the explanation of the invention the rotating component will be described as being a rotor surrounded by a casing. Filling the casing is a fluid such as air or water. In actual machines, the rotor may be a pump impeller as in the case of a centrifugal pump or compressor or it may be an armature in the case of an electric motor or generator.
When the rotor is rotating, a drag force is generated that acts against the direction of rotor rotation. The drag force that is generated is a direct consequence of the fact that the fluid mass in the machinery is simultaneously in contact with the moving rotor and the stationary casing for the machinery. This results in a drag force being transmitted through the fluid. Essentially, the rotor causes the fluid to rotate in the direction of rotation of the rotor due to the viscosity of the fluid. However, since the fluid is also in contact with the stationary outer casing the fluid motion is resisted by the casing. The result is that the rotor experiences a retarding drag force caused by the fluid in the casing. It should be noted that this fluid is present as an unavoidable consequence of the normal operation of the machine. For example, the fluid may be water (in the case of a centrifugal pump) or air (in the case of an electric motor).
The retarding drag force generated by the rotor rotation absorbs power. The amount of power absorbed primarily depends on the size of the rotor, speed of rotor rotation, the physical properties of the surrounding fluid (density and viscosity) and the clearance between the rotor and the surrounding casing. Well-designed rotating machines usually have the proper rotor shape and clearance between the casing and rotor in order to achieve reduce drag losses originating from fluid drag on the rotor.
Even with the proper rotor shape and clearance between the rotor casing, the rotor drag losses can be very high. An example will highlight the power losses than can be generated as a result of the retarding drag loss on a rotor. A centrifugal pump with an impeller 13 inches in diameter and rotating at 3600 rpm in water will experience a drag can exceed 20 hp. This is up to 30% of the power needed by the pump to operate.
In order to understand the operating principles of the invention, it is first necessary to understand the basic factors that determine the amount of fluid drag acting on a rotor. For simplicity, the rotor will be assumed to be shaped like a disk with a negligible thickness. There are three major factors that influence drag losses on a rotor operating in a given fluid. The first factor is the general shape of the rotor and casing, the surface roughness of the rotor and surrounding casing and the clearance between the rotor and casing. For well-designed rotating machines, the rotor and casing surfaces are smooth and unbroken and the clearance between the rotor and casing is relatively small. These design features all tend to reduce the amount of fluid drag acting against rotor rotation.
Another factor is the speed of rotation of the rotor. The drag force acting on the rotor is approximately proportional to the square of the speed of rotation. For example, doubling the speed of rotation increases the drag force by four times (i.e., 2.sup.2) and tripling the speed of rotation will increase the drag force by nine times (i.e., 3.sup.2). Likewise reducing the speed of rotation by one-half reduces the fluid drag drag acting on the rotor to one-fourth of the original value (i.e., 5.sup.2).
The third factor is the diameter of the rotor. For a given speed of rotation, the drag loss varies with the fifth power of the diameter. For example, doubling the diameter will increase the drag power loss by 32 times (i.e., 2.sup.5). Tripling the diameter increases the drag power loss by 243 times (i.e., 3.sup.5).
For many design reasons, a large diameter rotor is desirable. For example, in the case of a centrifugal pump, a large diameter rotor (or impeller as the rotor is normally called in a centrifugal pump) develops higher pressure than a small diameter rotor operating at the same rpm.
Designers for centrifugal pumps have had to resort to very elaborate schemes to reduce the fluid drag losses in the pumps. The most common way is to use multi-staging. Multi-staging is a technique of using two or more impellers of relatively small diameter. The fluid that is to be pumped passes sequentially through each impeller and each of the impellers provides a portion of the total pressure rise. Multi-stage pumps require very complex passages to route the fluid through the impellers and are therefore very expensive to manufacture. Such multi-stage pumps also contain many surfaces that are subject to high wear conditions and the pumps are therefore, expensive to maintain.
Another technique that has been used to reduce the fluid drag losses acting on the rotor, especially in centrifugal pumps, is to use a small impeller rotating at very high speeds (sometimes exceeding 10,000 rpm). This technique allows the use of a very small diameter impeller to achieve a given pressure increase thereby reducing the fluid drag losses acting on the impeller. This technique has two major drawbacks. One drawback is that a speed multiplier such as a gear box is frequently required between the motor that drives the pump and the pump to produce the high rotation rate. Such an item increases the purchase price and maintenance expense for the pump. The other major drawback is that high speed centrifugal pumps are prone to destructive fluid flow conditions such as cavitation due to the rapid acceleration experienced by the fluid inside the impeller.
The ideal pump for generating high pressure would have a single, large diameter impeller and would operate at relatively low speeds. As indicated above, however, a large diameter impeller creates a large fluid drag loss acting against rotor rotation.
A concept was developed in the early 1900's that theoretically allows fluid drag losses on a rotor to be drastically reduced by using freely rotating disks adjacent to the rotor. This concept is to place a freely rotating disk between the moving rotor and the stationary side wall of the rotor casing.
The amount of drag force acting on a rotor depends on how fast a rotor is revolving relative to the adjacent casing side wall. In typical rotating machinery, the casing side wall is stationary. However, a disk approximately with the same diameter as the rotor can be coaxially positioned between the rotor and the casing side wall. If this disk rotates in the same direction as the rotor the relative velocity between the rotor and the adjacent disk will be reduced resulting in a reduced fluid drag on the rotor.
The fluid mass adjacent to a rotating surface is dragged in the direction of rotation due to the fluid's inherent viscosity and the fluid mass rotates in the same direction as the rotor. The speed of the fluid rotation is approximately equal to the average of the speeds of rotation of the rotor and the casing sidewalls. For example, if the rotor is rotating at 3600 rpm and the casing sidewall is not rotating (which is typically the case), then the fluid mass will rotate at about 1800 rpm.
If a freely rotating disk is placed coaxially between the rotor and the stationary casing it can be shown that the disk rotates at about one-half of the speed of the rotor.
For example, if the rotor is rotating at 3600 rpm, the adjacent freely rotating disk will rotate at about 1800 rpm due to the combination of drag forces exerted by the fluid mass on each side of the disk. The fluid mass located between the freely rotating disk and the rotor revolves at about 2700 rpm which is the average of the 3600 rpm rotational speed of the rotor and the 1800 rpm rotational speed of the freely rotating disk. The fluid mass between the disk and the stationary casing side wall revolves at about 900 rpm which is the average of the 1800 rpm rotational speed of the disk and the 0 rotational speed of the stationary side wall. Note that without the freely rotating disk, the fluid mass rotated at 1800 rpm and with the freely rotating disk the fluid mass adjacent to the rotor rotates at 2700 rpm. This is a reduction of 900 rpm in the relative rotation rate or a reduction of 50%. As discussed earlier, since fluid drag loss obeys the square relationship, this 50% reduction in relative velocity reduces the drag loss by four times. In this example, if the fluid was water, the rotor and disks (one disk on each side of the rotor) are each 12 inches in diameter, then the rotor drag power loss would be about 3.5 horsepower. If the freely rotating disks were not present, then the rotor drag power loss would be about 14 horsepower. In this example the use of freely rotating disks reduces the power required to operate the pump by over 11 horsepower.
Although the concept of using freely rotating disks was suggested in the early 1900's there has been no practical application of this technology to reduce fluid drag losses. This technology has not been utilized for specific reasons due to the dynamics of the fluid pressure that are developed inside rotating machinery which will be more fully set forth below.
A rotating mass of fluid generates a radial pressure gradient having an intensity that is dependent upon the rotation rate of the fluid. Specifically, the pressure developed by the rotating fluid follows a square relationship. Doubling the speed of rotation increases the pressure gradient by 4 times. Tripling the rotation rate increases the pressure gradient by 9 times. For example, a mass of water 12 inches in diameter revolving at 3600 rpm would display a static pressure that is 240 lbs. per square inch higher at its periphery than at its center of rotation.
As indicated earlier, the fluid located between the freely rotating disks and the rotor rotates at about 3 times faster than the fluid rotating between the freely rotating disks and the stationary casing. Therefore, the rotating fluid mass between the disks and rotor will generate a pressure gradient about 9 times greater than the pressure gradient generated by the fluid between the disks and the casing side wall. For example, a 12 inch diameter rotor running at 3600 rpm in water with a 12 inch freely rotating disk will generate 135 lbs. per square inch pressure difference between the rotor center and the rotor periphery in the space between the disk and the rotor (here the fluid revolves at 2700 rpm). The fluid located between the disk and the stationary casing side wall will generate a pressure difference of about 15 lbs. per square inch between the center and the periphery of the disk (here the fluid revolves at 900 rpm). These unequal pressure gradients can create a very strong force on the freely rotating disk that acts in the axial direction. This axial force is a major problem that has prevented the commercial application of freely rotating disks.
In this case, the pressure at the center of the fluid mass located in the space between the disk and rotor will be 135 lbs. per square inch lower than the pressure at the outer periphery of the disk. Also, the pressure at the center of the fluid mass located in the space between the disk and the stationary side wall will be 15 lbs. per square inch lower than the pressure at the outer periphery of the disk. As is clear from the above, the fluid pressure is generally lower in the fluid mass between the disk and the rotor than the fluid pressure in the fluid mass between the disk and the stationary side wall. The generally higher pressure in the fluid mass between the disk and stationary side wall pushes the disk axially toward the rotor with a force that, in this example, exceeds 3,000 lbs.
On the other hand, if the pressure is equal on both sides of the disk at the disk center and there are seals at the outer periphery of the disk to allow a pressure difference to exist at the outer periphery, then the pressure at the outer periphery in the fluid mass located between the disk and the rotor will be 135 lbs. per square inch higher than the pressure existing at the center of the fluid mass. Also, the pressure at the outer periphery of the fluid mass located between the disk and the stationary side wall will be 15 lbs. per square inch higher than the pressure existing at the center of the disk. The generally higher pressure in the fluid mass between the disk and the rotor pushes the disk axially toward the stationary side wall with a force that, in this example, exceeds 3,000 lbs.
From the above, it is clear that a powerful axial force caused by the differing pressure gradients on either side of the disk will act to force the disk either toward the rotor or toward the stationary side wall. The direction of this force depends on whether the pressures are equalized on both sides of the disk at the outer periphery or at the center of the disk.
In the past, it has been indicated that bearings can be used to maintain proper position of such a freely rotating disk. However, the ratio of axial force to allowable bearing drag force can easily exceed one thousand to one (for example, the bearing drag should not exceed 3 lbs. of force yet that same bearing must be able to handle 3,000 lbs. of axial force). Friction thrust bearings can not reliably achieve this level of performance. Anti-friction thrust bearings (e.g. roller bearings) are very expensive in the required sizes (bore size can frequently be larger than 12 inches), must operate in the fluid surrounding the rotor (which is typically non-lubricating and may be corrosive, abrasive and/or hot), and must take an absolute minimum of room. Anti-friction thrust bearings would also create high maintenance costs and introduce unfamiliar field maintenance procedures. Thus, the use of bearings has not been an adequate solution to position the freely rotating disk to handle the axial forces that are generated on either side of the disk.
The present invention utilizes a structure that can cancel out the axial forces acting on the freely rotating disk thereby eliminating the need for large and expensive bearings capable of handling high axial thrust loads. The invention utilizes seals that can be located on each side of the idler disk and a pressure equalization port to completely cancel out the axial forces. These components work in a manner that when an imbalance in the axial forces cause the freely rotating disk to move from its desired operating position then the seals and equalization ports create a restoring force that acts in the opposite direction to the imbalancing force thereby restoring the freely rotating disk to the desired operating position. The restoring force is generated in a fraction of a second and acts to move the freely rotating disk either away from the rotor or away from the side walls of the outer casing as required.